Assessing the contribution of tumor mutational phenotypes to cancer progression risk.

Cancer occurs via an accumulation of somatic genomic alterations in a process of clonal evolution. There has been intensive study of potential causal mutations driving cancer development and progression. However, much recent evidence suggests that tumor evolution is normally driven by a variety of mechanisms of somatic hypermutability, which act in different combinations or degrees in different cancers. These variations in mutability phenotypes are predictive of progression outcomes independent of the specific mutations they have produced to date. Here we explore the question of how and to what degree these differences in mutational phenotypes act in a cancer to predict its future progression. We develop a computational paradigm using evolutionary tree inference (tumor phylogeny) algorithms to derive features quantifying single-tumor mutational phenotypes, followed by a machine learning framework to identify key features predictive of progression. Analyses of breast invasive carcinoma and lung carcinoma demonstrate that a large fraction of the risk of future clinical outcomes of cancer progression-overall survival and disease-free survival-can be explained solely from mutational phenotype features derived from the phylogenetic analysis. We further show that mutational phenotypes have additional predictive power even after accounting for traditional clinical and driver gene-centric genomic predictors of progression. These results confirm the importance of mutational phenotypes in contributing to cancer progression risk and suggest strategies for enhancing the predictive power of conventional clinical data or driver-centric biomarkers.

Toward Recovering Allele-specific Cancer Genome Graphs.

Integrated analysis of structural variants (SVs) and copy number alterations in aneuploid cancer genomes is key to understanding tumor genome complexity. A recently developed algorithm, Weaver, can estimate, for the first time, allele-specific copy number of SVs and their interconnectivity in aneuploid cancer genomes. However, one major limitation is that not all SVs identified by Weaver are phased. In this article, we develop a general convex programming framework that predicts the interconnectivity of unphased SVs with possibly noisy allele-specific copy number estimations as input. We demonstrated through applications to both simulated data and HeLa whole-genome sequencing data that our method is robust to the noise in the input copy numbers and can predict SV phasings with high specificity. We found that our method can make consistent predictions with Weaver even if a large proportion of the input variants are unphased. We also applied our method to The Cancer Genome Atlas (TCGA) ovarian cancer whole-genome sequencing samples to phase SVs left unphased by Weaver. Our work provides an important new algorithmic framework for recovering more complete allele-specific cancer genome graphs.

Algorithms and Complexity Results for Genome Mapping Problems.

Genome mapping algorithms aim at computing an ordering of a set of genomic markers based on local ordering information such as adjacencies and intervals of markers. In most genome mapping models, markers are assumed to occur uniquely in the resulting map. We introduce algorithmic questions that consider repeats, i.e., markers that can have several occurrences in the resulting map. We show that, provided with an upper bound on the copy number of repeated markers and with intervals that span full repeat copies, called repeat spanning intervals, the problem of deciding if a set of adjacencies and repeat spanning intervals admits a genome representation is tractable if the target genome can contain linear and/or circular chromosomal fragments. We also show that extracting a maximum cardinality or weight subset of repeat spanning intervals given a set of adjacencies that admits a genome realization is NP-hard but fixed-parameter tractable in the maximum copy number and the number of adjacent repeats, and tractable if intervals contain a single repeated marker.

Reconstructing ancestral gene orders with duplications guided by synteny level genome reconstruction.

BACKGROUND: Reconstructing ancestral gene orders in the presence of duplications is important for a better understanding of genome evolution. Current methods for ancestral reconstruction are limited by either computational constraints or the availability of reliable gene trees, and often ignore duplications altogether. Recently, methods that consider duplications in ancestral reconstructions have been developed, but the quality of reconstruction, counted as the number of contiguous ancestral regions found, decreases rapidly with the number of duplicated genes, complicating the application of such approaches to mammalian genomes. However, such high fragmentation is not encountered when reconstructing mammalian genomes at the synteny-block level, although the relative positions of genes in such reconstruction cannot be recovered. RESULTS: We propose a new heuristic method, MULTIRES, to reconstruct ancestral gene orders with duplications guided by homologous synteny blocks for a set of related descendant genomes. The method uses a synteny-level reconstruction to break the gene-order problem into several subproblems, which are then combined in order to disambiguate duplicated genes. We applied this method to both simulated and real data. Our results showed that MULTIRES outperforms other methods in terms of gene content, gene adjacency, and common interval recovery. CONCLUSIONS: This work demonstrates that the inclusion of synteny-level information can help us obtain better gene-level reconstructions. Our algorithm provides a basic toolbox for reconstructing ancestral gene orders with duplications. The source code of MULTIRES is available on https://github.com/ma-compbio/MultiRes .

FPSAC: fast phylogenetic scaffolding of ancient contigs.

MOTIVATIONS: Recent progress in ancient DNA sequencing technologies and protocols has lead to the sequencing of whole ancient bacterial genomes, as illustrated by the recent sequence of the Yersinia pestis strain that caused the Black Death pandemic. However, sequencing ancient genomes raises specific problems, because of the decay and fragmentation of ancient DNA among others, making the scaffolding of ancient contigs challenging. RESULTS: We show that computational paleogenomics methods aimed at reconstructing the organization of ancestral genomes from the comparison of extant genomes can be adapted to correct, order and orient ancient bacterial contigs. We describe the method FPSAC (fast phylogenetic scaffolding of ancient contigs) and apply it on a set of 2134 ancient contigs assembled from the recently sequenced Black Death agent genome. We obtain a unique scaffold for the whole chromosome of this ancient genome that allows to gain precise insights into the structural evolution of the Yersinia clade.

ANGES: reconstructing ANcestral GEnomeS maps.

SUMMARY: ANGES is a suite of Python programs that allows reconstructing ancestral genome maps from the comparison of the organization of extant-related genomes. ANGES can reconstruct ancestral genome maps for multichromosomal linear genomes and unichromosomal circular genomes. It implements methods inspired from techniques developed to compute physical maps of extant genomes. Examples of cereal, amniote, yeast or bacteria ancestral genomes are provided, computed with ANGES. AVAILABILITY: ANGES is freely available for download at http://paleogenomics.irmacs.sfu.ca/ANGES/. Documentation and examples are available together with the package. CONTACT: cedric.chauve@sfu.ca.